Chemoprevention studies in lung cancer have not yet resulted in any reduction in the lung cancer mortality.However, developments in biotechnology have made earlier diagnosis possible, and new classes of agents are developed targeting important pathways in the lung carcinogenesis. Thus, a new concept for chemoprevention studies are developed, which includes identification of high-risk groups with a high yield of preneoplastic changes. Traditionally, chemoprevention studies have used lung cancer as a primary endpoint, and thousands of patients have been included for a long study period. In order to get a higher throughput of chemoprevention studies including less number of participants, the selection of individuals with higher risk of developing lung cancer is important, as well as identification and validation of biomarkers for risk assessment and surrogate intermediate endpoints (intermediateeffect biomarkers). An intermediate-effect biomarker is a detectable lesion or biological parameter with some of the histological or biological features of preneoplasia or neoplasia but without evidence of invasion, which either is on the direct pathway from the initiation of the neoplastic process to the occurrence of invasive cancer, has a high probability of resulting in the development of cancer, or is a detectable biochemical abnormality which is highly correlated with the presence of such a lesion [l]. The intermediate-effect markers include detectable precancerous changes (confirmed by histology), an alteration of a gene considered to play a causative role and/or other indications of carcinogenesis, such as expression of a marker that represents the cause of a cancer. Intermediateeffect marker contribute to chemopreventive interventions in selection of target population, identification of subgroups of the population in whom precancerous lesions are detectable, in the measurement of outcome, in the interpretation of results, and in the characterization of molecular pathways involved. For lung cancer very few biomarkers have so far been validated. Atypical cells in sputum have shown to be predictive of lung cancer development. In the Colorado high-risk cohort moderateor more severe sputum atypia gave a significant increased hazard-ration (HR) of 1.6 compared to those individuals without sputum atypia. By adding sputum DNA methylation of 7 genes (p 16, MGMT, RASS FlA, H-Cadhedrin, DAP-kinase, PAX alpha, PAX beta) in a nested case-control study the HR increased to 10.2 [2]. Thus, sputum cell atypia seems to be useful to predict a high-risk population and even more when DNA methylation is added. Bronchial histology is considered a useful intermediate effect biomarker, although, there are very sparse data on the natural course of the bronchial changes. In some studies metaplasia index (fraction of biopsies with metaplasia) are used as an intermediate marker.In two studies a reduction in metaplasia index was found regardless of the chemoprevention treatment [3,4]. In one of the studies [3], smoking cessation during the trial showed a better correlation to metaplasia index than the treatment. However, the incidence in metaplasia varies from study to study. While in a study from MD Anderson a high prevalence of metaplasia occurred [5], in another study from the University of Colorado very few participants had metaplasia, but 55% had high-grade dysplasia [6]. In the latter institution several morphological models are currently validated as intermediate-effect endpoints including dysplasia index (the fraction of biopsies with dysplasia), the worst diagnostic score and the average score based on the total number of biopsies taken at each bronchoscopy.The scoring system is based on a scale from l-8 (1 = normal, 2 = hyperplasia, 3 = metaplasia, 4 = mild dysplasia, 5 = moderate dysplasia, 6 = severe dysplasia, 7 = CIS and 8 = invasive carcinoma) [7]. When using the histology as a surrogate endpoint the use of fluorescence bronchoscopy (LIFE bronchoscopy) has shown to be superior to the white-light bronchoscope for detection of preneoplasias [6]. The histological classification of bronchial changes has been validated by an international panel of pathologists, and a high reproducibility in the classification was observed [8]. Immunhistochemical markers such as Ki67, MCM2, ~53, EGFR, HER2 and apoptosis markers might be useful intermediate biomarkers, and results from ongoing validation studies are awaiting. Preliminary data has shown that increased Ki 67 index correlates to the increased level of dysplasia [9]. RAR-p has been shown to be modulated by 13-tic-retinoic acid treatment even in the absence of cytological improvement [ lo]. Strict validation programs are necessary in order to find clinical useful biomarkers for chemoprevention studies. Such a validation program should include several steps including intraand interlaboratorial reproducibility studies and nested case-control studies for the clinical validation. At the University of Colorado Cancer Center nested case-control studies are ongoing in order to find risk markers (i.e sputum atypia, DNA methylation and other biomarkers) in a high risk population (smoking history >30 pack-years and COPD). In chemoprevention studies bronchial histological changes and immunhistochemistry markers are validated as surrogate endpoints for treatment effect.